On Wed, 3 Dec 2003 14:44:18 +0000 (GMT), Philip Hazel
<ph10@???> wrote:
>On Wed, 3 Dec 2003, Marc Haber wrote:
>> Maybe an ACL to be evaluated after the reception was completed (even
>> after the acl_data) would be a good idea too.
>
>In a conventional configuration, where automatic delivery happens after
>a message is received, "after reception is complete" and "at the start
>of delivery" are almost the same thing (apart from re-deliveries).
Systems that do scanning are likely to have higher load, and are more
likely to trip queue_only_load settings.
Additionally, an ISP selling scanning services can't afford to ever
deliver any unscanned messages, and re-scanning on each queue run is a
performance killer on a large queue.
>Putting such code in the delivery function is much more logical, because
>that is where the code for bouncing at delivery time is.
I can understand that. However, it would be _very_ good to have some
means of "this filter has already been run successfully on this
message" in addition to "we have already started at least once to run
the filter on this message, but we don't know if the filter ever
finished".
Greetings
Marc
--
-------------------------------------- !! No courtesy copies, please !! -----
Marc Haber | " Questions are the | Mailadresse im Header
Karlsruhe, Germany | Beginning of Wisdom " | Fon: *49 721 966 32 15
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fax: *49 721 966 31 29
